[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: PAUP help?



----- Original Message -----
From: "David Peters" <davidrpeters@earthlink.net>
Sent: Saturday, August 04, 2007 5:25 PM

During a recent new analysis, I input matrix values in a typical
fashion. When I ran PAUP I got the following message and no results from
a heuristic search.

Missing data identified by '?'
Gaps identified by '-'

Warning: The 'Missing' option is not supported for PAUP*.  Distance
matrices are not allowed to contain missing data.

Deleting all question marks in favor of hyphens does not relieve the
problem.

Other files are fine. From experience or insight, does anyone have an
idea what might be the problem here?


...and later...

It was an unseen, still unknown bogie.
Redoing the character list fixed whatever the problem was.

No, it wasn't an unknown bug. You were trying to conduct a distance ( = phenetic) analysis, such as neighbor-joining or UPGMA, rather than a parsimony ( = phylogenetic) analysis. Enter the following command to change that back:


set criterion=parsimony

And don't use hyphens in morphological data matrices.

Contrary to what many people seem to believe, hyphens do _not_ mean "inapplicable" in PAUP*. PAUP* cannot handle inapplicable characters; you have to code them as "?".

Hyphens mean "gap in a molecular sequence" for PAUP*. Fortunately the default setting is that gaps are treated as missing data, which is an IMHO bad idea for molecular datasets, but ensures that accidental hyphens in morphological datasets are treated the same way as question marks, which is what you want. If you ever happen to have the other possible setting -- "gaps treated as fifth base" -- you'll probably end up with the presence of "-" being counted as a synapomorphy, which is nonsense from an evolutionary point of view (for morphological datasets) and must lead to totally spurious results.

Redoing the character list should not have changed anything, so it's a mystery for me that it did. However, you seem to have done a distance analysis. This is no good. Make sure "set criterion" is set to "parsimony" before you run a phylogenetic analysis.

If you still don't have the command reference, I can send it (it's a relatively small pdf file).